Agentic Builder - Claude Code

We're hiring builders to build production software with Claude Code on a per-project basis. You won't write much code by hand. You'll orchestrate the agents, design the skills, manage context and memory, and ship working apps. We need people who've actually mastered this and understand what's happening under the hood, not vibe-coders. If you've shipped real things with Claude Code and you can handle Git, deployments, databases, security, and the rest, we want to talk. You'll get a fixed pay per each project. Our average builders make more than $10k per month.

About Us

Bolder Apps is a product development studio that partners with US-based startups and established companies to build and scale innovative digital products. We specialize in AI-powered development, full-cycle product creation, and engineering team augmentation. Our mission is simple: build bolder, faster, and smarter.

Our Culture & Values (read before applying)

We move fast. We take ownership. We work with AI, not against it. And we expect everyone to bring ideas, not wait for instructions.

There are no daily checklists, no micromanagement, and no corporate politics. Instead, you’ll have autonomy, trust, and a team that’s always ready to help you grow. At Bolder Apps, impact matters more than titles, and curiosity matters more than seniority.

If you want a place where you can level up fast and actually see your work making a difference - welcome aboard.

Responsibilities

• Take projects from spec to deployed software, with Claude Code (and other coding agents where they fit) doing most of the typing
• Author CLAUDE.md files, skills, and prompt systems that produce consistent output
• Manage context, sub-agents, memory, and tokens across long-running tasks
• Drive TDD with agents using hooks, sub-agents, and verification patterns, and don't let the agent cheat
• Set up CI/CD, deployments, databases, secrets, and monitoring properly
• Be the final quality gate. Review what the agent produced, refactor what's wrong, ship what works

• Real, demonstrable Claude Code experience. You've shipped apps with it — not POCs, not toys. You can talk fluently about CLAUDE.md, skills, sub-agents, hooks, MCP servers, plan mode, and slash commands without consulting docs.
• Multi-agent fluency. You've worked across at least one other coding agent (Cursor, Codex, Copilot, Amp, Cline, or similar) and can articulate the trade-offs.
• Skill authoring. You've written reusable skills (SKILL.md, system prompts, agent definitions) that produce consistent output, not one-off prompts that need re-tuning every session.
• Context engineering. You manage the context window deliberately. You know when to summarize, when to spawn a sub-agent, when to clear and restart.
• Memory management. External memory, RAG, file-system memory, structured handoffs across sessions and days.
• Token economy. You optimize for cost. You know which model to use for which task. You're not burning Opus tokens on jobs Haiku would handle.
• Git. Branches, rebases, worktrees, conflict resolution, clean commits. You don't panic when something goes sideways with origin/main.
• Deployments. You can ship a Next.js app to Vercel, a backend to Fly or Railway, a container to AWS — and you can debug it when it breaks at 11pm.
• Databases. Postgres-class fluency. You've worked with Supabase or equivalent (Neon, RDS, PlanetScale). You understand schemas, indexes, migrations, foreign keys, transactions, and Row Level Security. You don't write SELECT * in production code.
• Security. Secrets management, env vars, auth flows (OAuth, JWTs, sessions), input validation, CORS, prompt injection in agent contexts, scoped tool permissions, the OWASP Top 10. You don't ship credentials to GitHub.
• Networking and infra basics. DNS, HTTPS, environment isolation (dev/staging/prod), basic observability (logs, error tracking, uptime monitoring).
• Reading code. When the agent produces something subtle, you catch it. You can debug across files, languages, and stacks.
• TDD with agents. You can make a coding agent actually do test-first development. You know it doesn't do this naturally — and you have your guardrails.
• Unit and integration testing. pytest, Jest, Vitest, Playwright, or whatever your stack demands. Meaningful coverage, not theater.
• Eval frameworks for agent quality, not just deterministic tests.
• Code review reflex. You don't merge what you haven't read.
• Honest communication. When something is harder than expected, you flag it early. You don't disappear and resurface with surprises.

• $10,000+ per month (paid per project). Uncapped upside for top operators who can ship multiple projects.
• Per-project structure. Defined scope, defined deliverables, no fake urgency.
• Fully remote, async-friendly. Work where and when you do your best work.
• Real autonomy. You pick the agent architecture, the skills, the workflow. We don't micromanage process.
• Steady pipeline. For engineers who consistently deliver, repeat work is the default.
• Direct line to decision-makers. No PM layers between you and the people who own the outcome.
• Tooling budget for model APIs, plugins, and infrastructure required to do the job well.
• A peer network of other top-tier agentic operators we work with on overlapping projects.

How to Apply

• A real project you shipped where AI agents wrote most of the code under your direction. Live URL preferred. Repo if it's open. Tell us what the agent got right, what it got wrong, and what you did about it.
• A skill, CLAUDE.md, MCP server, or hook setup you're proud of. Show us your taste and your standards.
• Your TDD-with-agents approach. Briefly: how do you make a coding agent actually do test-first development without cheating?
• Something that proves you operate under the hood. A war story about a deployment that went sideways. A Supabase schema you're proud of. A security issue you caught. A migration you executed without drama. Something that shows you're not just driving the agent — you understand the road.